Community Wishlist Survey 2022/Archive/Improve regex support in search engines and textual editor assistant

From Meta, a Wikimedia project coordination wiki

Improve regex support in search engines and textual editor assistant

NoN Proposes existing functionality

  • Problem: Regex support is poor both in the embedded search engine and in the textual editor assistant tool
  • Who would benefit: Everyone
  • Proposed solution: Provide full regex support, including the most sophisticated patterns
  • More comments:
  • Phabricator tickets:
  • Proposer: L736Etell me 17:45, 12 January 2022 (UTC)[reply]

Discussion

Hello @L736E:, thanks for this proposal. Would it be possible for you to elaborate on what would you like to see? What limits have you experienced? What would you like to do that is impossible now? SGrabarczuk (WMF) (talk) 04:16, 13 January 2022 (UTC)[reply]

  • I think the regex is like the title blacklist regex, but it's hard to understand. Maybe we should create a help page for advanced searches. Thingofme (talk) 11:05, 13 January 2022 (UTC)[reply]
  • @SGrabarczuk (WMF): A simple example: if I write in the search form say of en.wiki "([2-7]+2)nd Academy Awards" I would expect as most relevant results "22nd Academy Awards", "32nd Academy Awards" and on and the other less relevant not listed afterwards if not listed at all. Today I see that "2nd Academy Awards", that is "not" matching the regex used for searching, is displayed as "most relevant result". That's not correct or at least what an user would expect.--L736Etell me 11:13, 13 January 2022 (UTC)[reply]
    Thanks! I've instructed by a fellow Wikipedian whom I consider to be a regex expert that the search is a bit different from the regex feature in the editing mode. He says that in the case you've provided as an example, you need to type intitle:/([2-7]+2)nd Academy Awards/ or even intitle:/[2-7]+2nd Academy Awards/. Have you maybe taken a look at Help:CirrusSearch? SGrabarczuk (WMF) (talk) 17:51, 13 January 2022 (UTC)[reply]
    Re-pinging @L736E in case they missed the reply. MusikAnimal (WMF) (talk) 21:46, 20 January 2022 (UTC)[reply]
    @L736E Since we have not heard back from you, and what you're proposing appears to already exist, we're going to archive this proposal. Thanks for participating in the survey, MusikAnimal (WMF) (talk) 21:44, 26 January 2022 (UTC)[reply]
    Maybe I'm wrong, but that's because plain searches (Full text search in Cirrus help) doesn't support the use of regex at all (it uses its own set of regex-like characters).
    You need to use it with insource/intitle or with other modifiers. MarMi wiki (talk) 15:45, 22 January 2022 (UTC)[reply]